Description of problem: in part of scale up 4.1 testing we found lower rate of network sent recv bytes. I submit this as a bug in order to make sure that anything didn't broken. we need to clarify if some changes made around: - engine db queries results set. - engine db queries invocations (caching). - vdsm api calls results set - vdsm api invocations (caching) Version-Release number of selected component (if applicable): 4.0.1-0.3 How reproducible: 100% Steps to Reproduce: 1. scale up scenario of 1 host 1 SD and ramp up of 111 vms. 2. monitor the env for 12 hours. Actual results: lower network traffic were monitored Expected results: we expect to have better performance for build to build. currently we suspect that lower rate as a bug and not as enhancement. Additional info:
Closing. There's hardly anything we can do with this bug. There were tens, if not more of changes performed. For example, the 4.1 SDK is compressing by default, while 4.0 doesn't - did you use 4.1 SDK against 4.1 and 4.0 SDK against 4.0? (this is just an example, I believe you are using direct REST - which also, depending on the client, may or may not compress).